The numerical competence of rhesus monkeys will be investigated to provide an animal model of the non-linguistic representation of number. The proposed research will focus on the ability of monkeys to determine the ordinal and cardinal values of novel exemplars of familiar and unfamiliar numerosities. Ordinal competence will be assessed in experiments in which monkeys are first trained on the numerosities 1-4 in a list learning paradigm. They will subsequently be tested on pair-wise judgements of exemplar of the trained numerosities 1-4 and of the untrained numerosities 5-9. No reinforcement will be provided during tests with the novel numerosities 5-9. That will provide a basis for determining if monkeys have an ordinal representation of numerosity. Cardinal competence will be studied in a second series of experiments with a variation of the widely used matching-to-sample paradigm. Subjects will be required to match the sample by selecting a physically distinct choice stimulus that contains the same number of elements as the sample, i.e., to match solely on the basis of cardinal number. Reaction time data will be analyzed to model how monkeys represent ordinal and cardinal numerosities, and to evaluate the hitherto unsubstantiated claim that subitizing accounts for all non-human numerical discriminations.